Corneal versus refractive surgically induced astigmatism: vector analysis of their dissociation and implications for toric IOL planning

To determine whether tomographically measured surgically induced astigmatism (SIA) is reflected in the manifest refraction, and whether the ocular residual astigmatism (ORA) precedes surgery and is stable per eye.

Single-center, longitudinal, comparative vector analysis.

Eyes underwent non-toric intraocular lens (IOL) or phakic implantable collamer lens (ICL) implantation. Total corneal refractive power (Pentacam) and manifest subjective refractive astigmatism at the corneal plane were expressed as Cartesian double-angle power vectors before and 3 months after surgery, yielding tomographic SIA (tSIA), refractive SIA (rSIA), their vector difference ΔSIA, and the pre- and postoperative ORA. Each SIA was decomposed into its flattening effect and torque at the incision meridian. Centroids were tested against the null by eye side.

Of 290 eyes (256 IOL; 34 ICL), tSIA magnitude was small (IOL 0.47D; ICL 0.34-0.39D) with centroids near null. ΔSIA had a large mean magnitude (IOL 0.74-0.88D) but a near-origin centroid, significant only in IOL right eyes. Mean corneal flattening effect (+0.008D right, +0.140D left eyes) did not predict its refractive counterpart. ORA did not differ between cohorts (-0.11D, 95%-CI -0.25 to +0.02) and remained significant postoperatively, yet per eye only half the offset persisted in phakic and less in pseudophakic eyes.

The per-eye tSIA was uncorrelated with the refractive SIA and carried no systematic direction, as power vectors or flattening effect, which stayed below a quarter of a toric cylinder step. ORA preceded surgery and remained systematically stable. Individually measured tSIA has only limited value for toric IOL planning.